Nicolas explained the reverse is out of scope, and gave a reasonable explanation why (because much harder to reconstruct missing information IIRC). On Sun, Jun 13, 2021, 2:54 PM Denis Maier wrote: > Am 12.06.2021 um 11:39 schrieb Nicolas Goaziou: > > Hello, > > > > Denis Maier writes: > > > >> Yes, good this is coming. > > > > As a step forward, I rebased wip-cite-new branch with more support for > > note numbers handling. > > > > I added three customizable variables: > > > > - org-cite-adjust-note-numbers, which simply allows the user to toggle > > punctuation and note number moving (on by default). > > > > - org-cite-note-rules, which defines what rules to apply according to > > locale, expressed as a language tag, as in RFC 4646. > > > > - org-cite-punctuation-marks, which lists strings recognized as > > punctuation in the process. > > > > `csl' and `basic' processors now both make use of this. > > > > I'd appreciate some feedback, in particular about the docstrings of the > > variables above. I focused on the "note numbers" topic instead of > > "punctuation" since I found the latter too generic. > > > > Also, there are some points that may need to be discussed: > > > > - I'm not sure about the `org-cite-punctuation-marks' variable being > > global, i.e., not locale-specific. > > > > - There is no support for this in LaTeX-derived back-ends, because > > I don't know when a citation is going to become a footnote. As > > a reminder, there is no "\footcite" command in `biblatex' processor. > > OTOH, users might prefer using a more advanced mechanism, e.g., > > csquotes. > > > > - It doesn't do anything special in quote blocks, because I'm still not > > sure there is something to do. AFAIU, special casing there only > > applies to author-date location, which out of the scope of this code. > > > > WDYT? > > Ok, I've managed to test this a bit, and I think this looks pretty good > so far. > > The only question I'd still have is if this could somehow also cover the > reverse situation (going from a note style to author-date). I've noticed > that simply adding a new language rule doesn't work anymore---as opposed > to my initial tests with earlier iterations of that mechanism. Seems > like this mechanism is now only triggered when using a note based style. > > Best, > Denis > > > > > Regards, > > > >